the journal of chemistry vol. 269, no. 42, issue of 21, pp

THE JOURNAL OF BIOICGICAL. CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biolou, Inc

Vol. 269, No. 42, Issue of October 21, pp. 26479-26485.1994 Printed in U.S.A.

Features of Vacuolar H+-ATPase Revealed by Yeast Suppressor Mutants*

(Received for publication, July 27, 1994)

Frantisek Supek, Lubica Supekova, and Nathan Nelson4 From the Roche Institute of Molecular Biology, Roche Research Center, Nutley, New Jersey 07110

The yeast Saccharomyces cerevisiae serves as an ex- cellent model for the study of the structure and function of proteins. Numerous amino acid substitutions in the proteolipid subunit of yeast vacuolar H+-ATPase have been reported. Suppressed variants for several of the inactive mutants were selected after subjecting them to chemical or polymerase chain reaction mutagenesis and screening for second site suppressors. Suppressors for the mutation GlnW to Lys change were intragenic and resulted from the changes: Mal4 to Val, Val74 to Ile, Ilea' to Leu, and Ilea' to Met. These residues are found on three different transmembrane segments but presumably at the same surface of the membrane. A new inactive proteolipid mutation was constructed by changing Val'= to Leu. This residue is situated in the middle of the fourth transmembrane segment, neighboring Glu"' which is the potential dicyclohexylcarbodiimide-binding site. The intragenic suppressor mutations for the above amino acid replacement resulted in changes of Vals5 to Ala, Met68 to Val, or Ilelso to Thr. These residues are found in the second and fourth transmembrane segments, presumably on the same interface. It seems as if all those internal suppressor mutations compensate for the volume changes caused by the original displacement of the given amino acid. Five glycine residues, situated on the same face of the third transmembrane helix, were changed to valine and all these mutants were inactive. A suppressor mutation to one of those mutants (Gly"' to Val) was identified as substitution of ne1= to Val. The structural and functional implications of these findings are discussed.

Vacuolar H+-ATPase (V-ATPase)' energizes organelles of the vacuolar system in every known eukaryotic cell (14). The enzyme generates protonmotive force at the expense of ATP and functions as an ATP-dependent proton pump. The enzyme is composed of two distinct parts, a catalytic sector which functions in ATP hydrolysis and a membrane sector which functions in proton conduction across the membrane (2, 5-7). Each of these sectors is composed of several subunits necessary for the function and proper assembly of the enzyme (8-11). The prin- cipal subunit of the membrane sector is a highly hydrophobic protein (proteolipid) of about 16 kDa that binds DCCD. The binding of DCCD causes inhibition of proton pumping and ATPase activities of the enzyme and was implicated as a nec-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. Tel.: 201-235-3790;

The abbreviations used are: V-ATPase, vacuolar H+-ATPase; DCCD, dicyclohexylcarbodiimide; Mops, 4-morpholinepropanesulfonic acid; Mes, 4-morpholineethanesulfonic acid; PCR, polymerase chain reaction.

Fax: 201-235-5848.

essary component for proton transport across the membrane (12-17). Comparison of the amino acid sequences of the proteolipids from yeast and mammalian sources suggested that the proteolipid is the most conserved hydrophobic protein known so far (15, 18). The assembly and membrane topography of the proteolipid are not known. Since the proteolipid is inserted into the membrane without the aid of a cleavable signal sequence, it is not apparent to which side of the membrane its N and C termini are facing. Because of the sequence and functional similarities with the proteolipid of F-ATPases, it was proposed that the N and C termini of the proteolipid are facing the lumen side of the vacuolar system (4, 19).

Although plants and mammals may have two or three genes encoding proteolipids of V-ATPases, Drosophila and yeast con- tain a single gene encoding this protein (15,16,20,21). Several genes encoding V-ATPase subunits were cloned from yeast genomic libraries (9,15,20,22-24). With the exception of VPHl and STVl that encode homologous proteins, all the genes encoding other subunits are present as a single copy in the yeast genome (25). Disruption of each of these genes (except for VPHl or STVl ) gave an identical phenotype that cannot grow at a pH higher than 7 and is sensitive to low and high calcium concentrations in the medium (20, 26). This phenotype opened the door for site-directed mutagenesis studies. The assay for the function of the mutated genes is performed by transforming the interrupted mutants with plasmids containing intact or modi- fied genes. Growth at pH 7.5 indicates the presence of a functional gene. The proteolipid was the first subunit to be studied by site-directed mutagenesis (27). This study indicated that the proteolipid is quite sensitive to changes in its amino acids and even some hydrophobic residues that were changed into similar amino acids abolished the activity of the enzyme (27). The strict conservation of amino acid sequences in potential transmembrane helices suggests tight contacts among the transmembrane segments necessary for activity of the enzyme. As an initial step in understanding the membrane topography and interaction of the proteolipid with other subunits, we initiated a study of second site suppressors for an inactive mutant form of the yeast proteolipid.

EXPERIMENTAL. PROCEDURES East Strains and Analysis of East Mutants-Saccharomyces cerevi-

siae strain W303-1B was used throughout this study. The haploid strain was MATa, leu2, his3, ade2, t r p l , w a d , and the corresponding disruptant mutants were NNYll-uma3:;LEU2, NNY12-Auma3::URA3, NNY72-tfpl::LEU2, NNY22-vma2::LEU2, NNY45-uma5::LEU2, in which the genes encoding subunits c (proteolipid), A, B, and C were interrupted (20, 24, 29). The cells were grown in a YPD medium containing 1% yeast extract, 2% bactopeptone, and 2% dextrose. The medium was buffered by 50 mM Mes and 50 mM Mops, and the pH was adjusted by NaOH (20, 27). Agar plates were prepared by the addition of 2% agar to the YPD buffer medium at the given pH. Yeast transformation was performed as described previously (28), and the transformed cells were grown on minimal plates containing a 0.67% yeast nitrogen base, 2% dextrose, 2% agar, and the appropriate nutritional requirements. In most experiments 0.1% casamino acids were added to

26479

26480 Second Site Suppressors in Yeast V-ATPase the minimal plates. Strain NNYl2 was generated by deleting most of the gene encoding the proteolipid from Leuso to its C-terminal amino acid. About 0.45 kilobase pair of the 5' end of the gene including the reading frame up to Leus0 and 0.3 kilobase pair of the 3' end of the gene starting at the last amino acid were cloned by PCR to include BamHI a t the 5' end, SphI at the 3' end, and EcoRI and Sal1 sites between the two fragments. The DNAfragment was cloned into BamHI and SphI sites of the plasmid YPNl (27) and then URA3 was cloned into the EcoRI and Sal1 sites generated by the PCR. The resulting BamHI-SphI fragment was used for transforming W303-1B yeast strain. Colonies that grew on minimal agar plates without the addition of uracil were analyzed by PCR and the deletion mutation of proteolipid (NNY12) was verified by DNA sequencing. The phenotype of the various V-ATPase mutants was analyzed by growing the cells on YPD plates buffered at pH 5.5 or pH 7.5 as described previously (20, 27). Similarly, the proteolipid mutants in which single amino acids were exchanged were assayed by transforming the proteolipid disruption or deletion mutations (NNY11 or NNY12) with the mutated proteolipid gene and following the growth of transformants at pH 5.5 and 7.5 (27). Lack of growth on agar plates buffered a t pH 7.5 indicates inactive V-ATPase. Vacuoles isolated from inactive mutants exhibited less than 5 6 proton pumping activity in comparison with vacuoles from wild-type cells.

Chemical Mutagenesis and Assay of Second Site Suppressor Mutants-A mutant bearing a uma3 disruption or deletion mutation was transformed by a YPNl plasmid containing the proteolipid gene in which LYS?~ was substituted by Asn and GlnS0 was substituted by Lys. The transformed cells were grown on minimal agar plates without tryptophan. Individual colonies were analyzed for lack of growth a t pH 7.5 (27) and inoculated into a liquid minimal medium lacking tryptophan. The cells were harvested, suspended in water at a cell density of 2 optical density units a t 600 nm, and treated with 5 pVm1 ethyl meth- anesulfonate for 60 min. The treated cells were washed with 5 6 sodium thiosulfate, followed by washing in water, and resuspended in a minimal medium without tryptophan and growing at 30 "C overnight. This step eliminated mutants that are unable to grow in these conditions. The cells were harvested by centrifugation a t 5,000 x g, suspended in water at a cell density of 10 optical density units a t 600 nm and plated on YPD buffered a t pH 7.5. About 100 independent colonies were collected and analyzed for the site of the second mutation as follows. The YPNl plasmid was isolated from each yeast colony by cloning into Escherichia coli cells using published procedures (30). The isolated plasmids were used for transforming proteolipid disruptant mutants. Isolated colonies were assayed for growth a t pH 7.5, and if positive the entire gene encoding the proteolipid was sequenced. Yeast colonies maintaining the original mutated proteolipid were cured from the plasmid by growing them in a YPD (pH 5.5) medium and then screened for colonies that will not grow on minimal medium without tryptophan and onYPD (pH 7.5). Then they were transformed with the original mutated proteolipid. Growth on buffered YPD a t pH 7.5 is an indication that the second suppressor mutation is situated in the yeast genome.

Generation of Suppressor Mutations by PCR-Mutations giving changes in the vicinity of the DCCD-binding site GIu'~' in the proteolipid were generated as described previously (27). Second site suppressor mutations for these mutations were generated by PCR as follows. The mutated proteolipid gene, with the substitutions Val138 to Leu or IleI3' to Ala or Gly'" to Val, were amplified under limited amounts of each of the triphosphodeoxynucleotides. Four reaction mixtures containing a 200 PM concentration of three of the dNTP's and a 2 p~ concentration of the fourth dNTP were amplified by four PCR cycles using the original proteolipid primers containing EcoRI at the 5' and SphI at the 3' end (27). Then the four reaction mixtures were mixed and run through 30 PCR cycles. The mutations are generated during the four cycles in which one of the dNTP's is present in limited amount and amplified in the following cycles in which all dNTP's are present in saturated amounts. The resulting DNA fragments were cloned in the EcoRI-SphI sites of a YPNl plasmid, and following amplification in E. coli cells the mutated plasmid library was used for transforming the disruptant yeast mutants NNY12, in which most of the proteolipid gene was deleted (see above). The transformed cells were grown on minimal plates lacking uracil and tryptophan and about lo5 independent colonies were collected and kept frozen at -80 "C in the presence of 15% glycerol. For the selection of suppressor mutants the cells were thawed on ice and spread on 20 YPD plates (15 cm) buffered at pH 7.5. The colonies that grew under this condition were collected separately, their plasmids were isolated, reexamined for activity of the proteolipid gene, and subjected to DNA sequencing of the entire gene.

Preparation of Yeast Vacuoles-The NNYll strain transformed with theYCp5O plasmid containing the proper variant of the proteolipid gene

2% GLUCOSE (pH5.5) 0.3% LACTATE (pH5.5 )

3% LACTATE (pH5.5 )

3% GLYCEROL (pH 5.9 3% PYRUVATE (pH 5.5)

FIG. 1. Null mutants inV-ATPase have no Pet- phenotype. Yeast cells were grown on plates containing 1% yeast extract, 2% bactopeptone, 50 mxl Mes, 50 mM Mops, 2% agar, and the indicated carbon source. pH was adjusted to the indicated value by NaOH. Top, wild-type cells W303B; bottom, p- cells; left, NNYll cells; right, NNYl2 cells.

was grown in minimal medium lacking uracil and used for the inocu- lation of a 5-liter YPD medium buffered to pH 5.5. After six generations (AGm = 0.8-1.0) cells were collected, and vacuolar membranes were isolated as was described (31). ATP-dependent proton uptake activity was assayed according to published procedure (32).

RESULTS

The Phenotype of V-ATPase Disruptant Mutants-The presence of an active V-ATPase is necessary for viability of most eukaryotic cells (2-4). Energization of the vacuolar system by this enzyme drives several vital transport processes across membranes of the various organelles derived from the vacuolar system (33). Fortunately, in yeast we could obtain a viable phenotype when the genes encoding various V-ATPase subunits were interrupted (20). The mutants overcame the lack of acidification of their vacuoles by endocytosis of external fluid.' Therefore, the mutants can be grown only a t low pH (20). How- ever, they cannot overcome high and low calcium concentrations in the medium or utilize nonfermentable carbon sources (27, 34-37). As shown in Fig. 1, the proteolipid disruptant mutants in strain W303 are able to grow on 3% glycerol. In contrast, they did not grow on other nonfermentable carbon sources such as lactate and pyruvate when present at a concentration of 3% in the medium. However, when the lactate concentration was reduced to 0.3%, the mutant cells grew well. A similar effect was observed with pyruvate as a carbon source (not shown). The same results were obtained with yeast mutants in which the genes encoding subunits A, B, C, and E were disrupted. The sensitivity to 3% lactate or pyruvate did not

* H. Riezman, personal communication.

A 40

I G V

L R D

G S

D L

K L 50

A F

G T

y~ 30

A G A

L T S F

I

C P

M TEL

B

K N

I v v

'MI A 60 G

I I I A

Y L G v s L

70 v v

s y G L

80 0 K 0

A L Y

40 ,c A T C "

G' V

G S K

A

G T

Ay 3o A G

S F L T

I I A 20

A; GC

; F G P 10

V Y A

C P

M TEL

Second Site Suppressors in Yeast V-ATPase 26481

TABLE I Amino acid changes that suppress inactive mutations

in the yeast proteolipzd Change caused by Original change suppressor mutation Growth at pH 7.5

- 120 1) Lys34 to ksn

s s Q Q P R , 2) ~ 1 ~ 9 0 to L ~ S - Mal4 to Val +

R G -

GV A D

G I 110 V G

I A F

G A

AGL 1w s L

V L G G S

L R P D L L 50

F

N K

A-yl V

V w M G A M,

LG

P I

I I I A

70 v V v s L V

C Y

S

G L

800 K

Leu73 to Phe IleI3O to Ser

- ~~~ ~ ~

F VaV4 to Ile

Iless to Met V

M L G I 130

11e89 to ~ e u

I I L 3) Mal4 to Val F E 4) VaV4 to Ile

5 ) Leu'31 to Ala L 6) to Ala

Y L 7) Leu'33 to Ala

V L A 8) Ile'34 to Ala

R vv

A V L G 140

G I

150

N 160

S C 9) Phe'35 to Ala 10) Ala136 to Val

12) to Ala 13) Val'38 to Leu

A T Q D 11) Ala'36 to Leu

Vals5 to Ala Met6' to Val Ile'30 to Thr

14) Gly% to Val 15) GlY8 to Val 16) Gly'O' to Val

Ile'34 to Val

+ + + + + + 4-

- + +

+!- + + + - 4- + 4- - - - + -

120 17) Glyio5 to Val 18) Gly'O' to Val -

s s Q Q P R -G L R F

V chemically mutagenized and screened for suppressors as de- G

A D G v scribed under "Experimental Procedures." This procedure

V G G I 7'0 allows the generation of both extra and intragenic suppressors.

A F I

G A L G 140 tragenic and were not studied further. All second site mutations A V-L F I E All suppressors to the mutation Lys34 to Asn change were ex-

AGL

G S V A L v G L

L

G I 1CQS L Y L that suppressed the mutation causing the Glngo to Lys change

N IM) a predicted membrane organization of the proteolipid according L S C t o the hydropathy plot, in which the residues changed by the

R vv

were intragenic. They included the changes Ile8' to Leu or Met, Val74 to Ile, to Val, and GlnW to Asn (Table 1). Fig. 2 shows LIS0

G A

O W I A T Q D suppressor mutations can be seen to be on the same side of the

- F membrane as the residue altered by the original mutation and 0 A ~ ~ ~ u at the same face of the membrane where the N and C termini

according to ita hydropathy plot. A, the amino acid alteration Ging0 FIG. 2. Organization of the proteolipid in the membrane

to Lys (depicted in bold letters) is suppressed by second site mutations

Mal4 to Val or Val74 to Ala. Replaced amino acids are underlined. B, that result in changes of (indicated by arrows) Ile*' to Leu, Iles9 to Met,

original replacement Val'38 to Leu (marked with bold letters) is suppressed by mutations that result in changes of Ile'33 to Thr, VaP6 to Ala or Met59 to Val (positions where the changes occurred are underlined).

result from the high ionic stress, because the addition of equivalent amounts of NaCl to media containing 0.3% lactate or pyruvate did not prevent growth of the mutant cells. However, it was reported that V-ATPase mutants in other strains are unable to grow on glycerol and other nonfermentable carbon sources and therefore suggested Pet- phenotype for these mutants (11,26,38). Because glycerol is considered to be a classi- cal nonfermentable carbon source (39,401, and our mutants are able to grow on it as the sole carbon source, the V-ATPase mutants cannot be classified as having a Pet- phenotype.

Second Site Suppressor Mutants in the Proteolipid Subunit of the Membrane Sector-Previously we reported on the effect of 65 amino acid changes in the yeast proteolipid on growth at high pH (27). Only 24 of those substitutions resulted in inactive proteolipid that could not support growth at pH 7.5. ' Ibo of the inactive mutants, Lys34 to Asn and Glngo to Lys (see Fig. 2 for the position of the correspondent amino acid residues) were

reside. Figs. 3 and 4 depict experiments measuring the proton up-

take capability of the suppressed mutants in comparison with wild-type and inactive mutants. Although the original mutants containing the amino acid change GlngO to Lys failed to accu- mulate quinacrine, all suppressed mutants accumulated quinacrine into their vacuoles (Fig. 3). This observation suggests an active proton uptake into the vacuoles of the cells derived from the suppressed mutants. Fig. 4 shows the ATP-dependent proton uptake of vacuoles isolated from the above mentioned strains. The vacuoles of the mutant Glngo to Lys show no activity of ATP-dependent proton uptake. All suppressed mutants showed restoration of this activity to various degrees. One of them, bearing an Ilea' t o Leu change as well as the original change, was even more active than the wild-type. The Mal4 to Val change resulted in only about 20% of the wild-type activity. This differential activity is not apparent from the quinacrine accumulation experiment shown in Fig. 3. Therefore, even par- tial restoration of proton pumping activity is sufficient for res- cuing the wild-type phenotype.

Proteolipids of F- and V-ATPases are the DCCD-binding sites of the corresponding enzymes (12-18,4144). Due to sequence homology it was suggested that in V-ATPases G ~ u ' ~ ~ is the DCCD-binding site, and this residue may be involved in proton conductance. Substitutions of G I u ' ~ ~ to Gln, Val, or Lys inacti-

26482 Nomarsk i

Second Site Suppressors in Yeast V-ATPase

Fluorescence +MgATP I

FIG. 3. Vacuole acidification by V-ATPase containing GlnW to Lys m u t a n t proteolipid and its suppressed derivatives. NNYll strain transformed with YPNl plasmid carrying the indicated proteolipid variant gene was grown in minimal medium lacking tryptophan to exponential phase. 1 ml of culture was spun down, resuspended in 0.1 ml of YPD containing 100 mM HEPES (pH 7.6), and 200 p~ quinacrine and incubated for 10 min a t 30 "C. After three washes with ice-cold 100 mM HEPES-NaOH (pH 7.6), 2% glucose buffer, cells were finally resuspended in 0.1 ml of the same buffer. Resuspended cells were mixed with equal volume of 1% low melting point agarose, mounted on the glass slides, covered with a coverslip, and observed within 10 min. Vacuoles were visualized using Nomarski ( lef t ) and fluorescence (right) microscopy. Proteolipid indicates proteolipid disrupted mutant transformed by the plasmid YPNl containing the gene encoding the wild-type proteolipid.

vated the proteolipid, and even substitution to Asp was barely active (27). Amino acid residues situated in close proximity to AspG1 in the E. coli proteolipid were shown to be more sensitive to changes than those situated in other parts of the protein (4144). Therefore, we generated substitutions in the amino acids surrounding Glu137 in the yeast proteolipid. As shown in Table I, substituting the amino acids at positions 131, 132, 133, and 138 to alanine resulted in an active proteolipid. Substitu- tion of Phe135 to Ala gave rise to a slow growth at pH 7.5 and substituting Ile134 to Ala resulted in an inactive mutant. The latter was compensated by two independent second site suppressors Leu73 to Phe and Ile13' to Ser.

Although the substitution Ala136 to Leu had no effect on the

1 rnin ' Time ' +FdCP

FIG. 4. Proton uptake of GlnW to Lys mutant is rescued by suppressor mutations. About 15 pg of vacuolar membranes were added to 1 ml of medium containing 20 mM Mops-Tris (pH 7.0), 0.15 M KCl, and 15 1.1~ acridine orange. Where indicated, 1 pmol of MgATP was added, and the internal acidification was monitored by the change in the A,, - A,,, difference. Where indicated, 1 nmol of carbonylcyanide p-(trifluoromethoxy)phenylhydrazone was added. Membranes were pu- rified from the NNYll strain transformed with YCp50 plasmid carrying the variants of the proteolipid gene encoding the indicated changes: 1, inactive mutant Glnw to Lys; 2, suppressed mutant Glnw to Lys + AlaI4 to Val; 3, Gln9" to Lys + HeR9 to Met; 4, GlnW to Lys + Val" to Ile; 5, wild-type proteolipid gene; 6, Gln"' to Lys + IleR9 to Leu.

growth at pH 7.5, similar substitution of Va113* to Leu resulted in an inactive proteolipid. The inactive Val138 to Leu was rescued by three second site suppressors of VaP5 to Ala, Met59 to Val or IleI3' to Thr. Fig. 5 shows the growth at pH 5.5 and 7.5 of wild-type, the null mutant (NNYll), the N N Y l l transformed by the mutant plasmid encoding the inactive Val13* to Leu change and its suppressor containing mutants. Both the null mutant and NNYll transformed with the plasmid encoding the inactive proteolipid could not grow on a medium buffered a t pH 7.5. On the other hand, the suppressed mutants grew well at this pH. As shown in Table I, the suppression resulted from amino acid changes inside transmembrane segments: to Ala and Met5' to Val occurred in the second transmembrane helix and Ile13' to Thr occurred in the fourth helix. It is note- worthy that the suppressor mutations resulted in replacements of large by smaller amino acids.

The third transmembrane helix is rich in glycines that are conserved in all the proteolipids of eukaryotic cells (18). These glycines are located a t one face of the helix and therefore may play a role in the assembly of the membrane sector or even in its mechanism of action. We substituted five of those glycine residues to valines and observed that all the mutations resulted in inactivation of the enzyme (Table I). A suppressor mutant for the substitution GlyIo1 to Val was isolated and identified as a second site substitution of Ile134 to Val in helix IV. This isoleucine residue is situated in close proximity to G I u ' ~ ~ which provides the DCCD-binding site and functions in proton conduction across the membrane.

Second Site Suppress0

pH 5.5 pH 7.5

to Leu can grow at pH 7.5. The NNYll strain was transformed by FIG. 5. Suppressed mutan t s of inactive proteolipid gene Val'%

YPNl plasmid containing a gene encoding the proteolipid, mutated proteolipid, or corresponding suppressor variant. Independent transformants were checked for growth on YPD medium buffered to pH 5.5 Cleft) and 7.5 (right). 1, wild-type proteolipid gene;2, nontransformed NNYll strain; 3, NNYll transformed with the inactive variant of the proteolipid gene encoding Val13R to Leu; 4, suppressed mutants bearing changes of Val'?' to Leu + Ile'33 to Thr; 5, Val'3n to Leu + Metsg to Val; 6, Val138 to Leu + V a P to Ala.

DISCUSSION The study of V-ATPase in yeast cells provide several unique

opportunities not available in any other system. In addition to their easy transformation, homologous recombination, and ge- netic information, so far they are the only eukaryotic cells that could sustain inactivation of their V-ATPase. I t was proposed that maintaining a protonmotive force across the membranes of the vacuolar system is vital for every eukaryotic cell, but yeast can grow a t low pH because they provide the acidic environ- ment to the interior of the vacuolar system by fluid-phase endocytosis (20). The protonmotive force is utilized for driving numerous secondary transport systems, among which calcium transport through the vacuolar membrane is necessary for maintaining calcium homeostasis in the cells. Although V- ATPase inactivated yeast cells can grow at low pH, they are sensitive to high and low calcium concentrations in the medium (11, 26, 27). Presumably the protonmotive force provided by fluid-phase endocytosis is not sufficient for a massive pumping of calcium leaked to the cytoplasm a t high external calcium concentrations, and it is not sufficient to release the calcium bound to EGTA. The apparent Pet- phenotype is more difficult to explain. How is energizing of the vacuolar system involved in res- piration? The simplest explanation would be that the vacuole stores some excess of respiratory metabolites and by so doing maintains a favorable balance among them. Such activity is known to take place in the photosynthesis of C-4 succulent plants (45). In these plants malic acid is accumulated at night, and this accumulation is dependent on the activity ofV-ATPase. Likewise, the yeast vacuole may function in V-ATPase-dependent uptake of excess metabolites. This would explain the growth of the yeast mutants on a medium containing low concentrations of nonfermentable carbon sources. Transport systems that may be involved in these processes are under investigation.

The fine structure of V-ATPases is not known. Crystalliza- tion attempts are hampered by lack of a convenient and abun- dant source of the enzyme and by the property that the catalytic sector by itself has no ATPase activity (8, 46). Therefore, we initiated a project of site-directed mutagenesis to better understand the subunit structure of V-ATPases (24, 26). Sub- units A and B of the catalytic sector and the proteolipid of the membrane sector are the most conserved subunits of the enzyme. Accordingly they are more sensitive to changes in their amino acid sequences than the remaining subunit^.^ Even

L. Taiz, S. L. Taiz, H. Nelson, and N. Nelson, manuscript in preparation.

Irs in Yeast V-ATPase 26483

though every subunit of the V-ATPase is essential for its activity, there is a big difference in their sensitivity to mutations. Although subunit C of the catalytic sector can tolerate numerous changes in its amino acid composition, the proteolipid, a subunit of the membrane sector, is quite sensitive to even con- servative substitutions. Therefore second site suppressors were utilized for the analysis of its structure and function.

The proteolipid is among the most conserved hydrophobic proteins in nature (18). It is quite sensitive to subtle amino acid changes, even at the transmembrane helices (27). We selected second site suppressor mutants for some of the original innctivc mutations. All suppressors for the mutation encoding the L ~ s ' ~ to Asn change were extragenic, suggesting interaction with other subunits (48). On the other hand, all suppressors for the mutation encoding the GlngO to Lys change were intragenic. Amino acid GlnS0 is probably located at the opposite side of the membrane to L Y S ~ ~ . The lack of extragenic suppressors for the GlngO to Lys mutation may indicate that this side of the proteolipid has little or no interaction with other subunits. If cor- rect, it may imply that is facing the catalytic sector on the cytoplasmic side of the membrane and GlngO is in the lumen. This organization is in line with previous proposals derived from the amino acid homologies with the proteolipids of F-ATPases (4, 18, 19). Lysine residue is positively charged; nevertheless, all suppressors for GlngO to Lys mutation resulted from exchanging neutral amino acids. Glutamine is smaller than lysine but its amide group is more bulky than the corresponding methylamine part of lysine. All amino acids that suppress the mutation GlngO to Lys are either larger in size or have a changed topology of the side chain when compared with the original residue. This suggests that suppression of the original mutation exhibited space filling properties and the residues Ala14, Val74, Ilea', and GlngO are located close to each other in a tightly packed structure. The approach of replacing charged amino acids with neutral residues inside the transmembrane helices, together with isolation suppressors, was successfully employed for the identification of potential contacts between the helices. (41,491. Because of the lack of charged amino acids in the transmembrane part of the proteolipid (with the exception of G ~ u ' ~ ~ ) and obvious sensitivity to changes in the size of side chains, we attempted to devise an alternative approach for the mapping of helical contacts. The proteolipid was first mutagenized and residues Ala'36 and Val'38 were replaced with either larger (Ala'36 to Val, Ala'36 to Leu, Val'3a to Leu) or smaller (Val'38 to Ala) amino acids. In the next step, suppressed variants of the mutant bearing the inactivating Val'38 to Leu change were isolated. Suppression resulted from exchanges either at Ile130 to Thr or Val5s to Ala or Met5' to Val. Fig. 6 depicts a computer simulation of the area of helices I1 and IV, where the change caused by the original and suppressor mutations occur. Apparently, all second site mutations are replacements of larger amino acids by smaller ones, thus compensating for the original change in the inactive mutant. Even though this result is not possible to interpret in simple mechanistic terms, the character of changes as well as the possibility of arranging helices I1 and IV in such a way that 2 amino acids of helix IV (Val'38, Ile13') face Mets9 and Valss in helix 11, indicates potential contact between these two helices (Fig. 6). To further test this hypothesis, we constructed five additional mutations in helix W. Only the mutation encoding the change Ile'34 to Ala resulted in inactive proteolipid. This residue is present in close proximity to G ~ u ' ~ ~ and a suppressor to this mutation was identified on the same face of the helix at the position of Ile'". Sensitive positions were identified in helices I1 and IV, where inactive substitutions and suppressor mutations were localized. Among them, changes in residues Ile'34 and Val138 resulted in inactive

26484 Second Site Suppressors in East V-ATPase IV

A

1130

C

L1

E

IV II

/ B

v55

D

F

L1

possible contact of helices I1 and lV. The pictures show a-helical FIG. 6. Suppressors of inactive mutation Val" to Leu indicate

wheels and amino acids in positions 55, 59, 130, and 138 with space- filled hydrogen atoms of side chains. Hydrogen atoms of the original amino acids are depicted as shaded circles, leucine replacing valine in inactive mutant is contoured with a thick black line, and sup- pressing amino acids are drawn with black circles. Assignment of the a-helical structure is based on hydropathy plot analysis (27). A, amino acids MeP, Ile13', and Val'38 are arranged in close proximity in helical wheels of transmembrane segments I1 and IV (top view). B, as in "A" but side view. D, inactive mutation Val138 to Leu introduces larger amino acid. C , the mutation in "D" suppressed by M e P to Val or 113" to T (E) or VS6 to A (F) . Pictures were drawn by using the program MacImdad Version 4.0.

proteolipid, and suppressor mutations were found in Val55, Met59, LeuT3, and IleI3O (Table I). This sensitivity is consistent with a possible interaction between these two transmembrane helices. A model, derived from analysis of high resolution elec- tron microscopy, suggested intermolecular interaction between helices I1 and IV in isolated proteolipid fkom Nephrops norue- gicus (52). It is apparent that in this stage of our study, we cannot differentiate inter- and intramolecular interactions between the helices of the proteolipid. Recent studies with mutations introduced into the hydrophobic core of chymotrypsin in- hibitor and interactions of a-helices in model system demonstrated that subtle changes in the size of the amino acid

residues in the interface of two helices can cause significant destabilization (47, 53). Our studies of site-directed mutagenesis and suppressor mutations are in line with these studies.

The proteolipid of V-ATPases is double the size of that of F-ATPases and evolved by gene duplication and fusion of the latter (14). It was suggested that this gene duplication confers most of the distinct properties of V-ATPases such as the exclu- sive function as a proton pump, the lack of proton conduction in membranes depleted of catalytic sectors, and operating below thermodynamic equilibrium (2, 46, 50). Since helices I1 and IV evolved from the same part of the original gene, it was inter- esting to observe that the suppressors of to Leu mutation were found on both helices. The structure of the proteolipid from E. coli F-ATPase was recently determined by NMR reso- nance assignments in organic solvents (42, 43). It folds across the membrane as a hairpin of two a-helices that are in close contact. Moreover, changes encoded by mutations that suppress the original mutation Alaz4 to Asp + AspG1 to Gly were localized on helix 11 of the proteolipid or on the transmembrane helices of subunit a (44). These suppressors appeared to en- hance the activity of the original mutation in which the nega- tively charged AspG1 at helix I1 was moved to the position of Alaz4 at helix I. The mutations described in this work were totally inactive and compensated by subtle changes in hydrophobic amino acids. Several attempts to move Glu13?, which is the homologous amino acid to AspG1 in the E. coli enzyme, yielded a n inactive enzyme that so far could not be suppressed. The equivalent of subunit a of F-ATPases in V-ATPases has as yet not been identified. Nevertheless, it is anticipated that a common mechanism of ATP-dependent proton uptake will un- derlie both F- and V-ATPases. The subtle changes of the suppressor mutants described in this work suggest a very tight conformation in the proteolipid of the active V-ATPase. This semicrystalline structure may occlude water molecules that act in the catalysis of energy coupling during ATP-dependent proton uptake activity of the enzyme.

The mechanism of proton conductance across membranes is poorly understood. A mechanochemical coupling between ATP hydrolysis in the catalytic sectors of F- and V-ATPases and proton transport across the membrane sector was proposed (see Ref. 51). Lack of sufficient charge groups and successful replacements of polar amino acids by glycine in the yeast proteolipid prompted us to suggest that water molecules coordinated to those glycine residues may take part in proton conductance (27). In helix I11 of the V-ATPase proteolipid there exist five glycines facing the same side of the helix (see Fig. 2). We mutated each of those glycines to valine residues and found that all these substitutions inactivated the proteolipid (Table I). This is in contrast to the overall inactivation frequency of about 35% observed in the previous extensive mutagenesis (27). Therefore, it is apparent that this face of helix I11 is very sensitive to volume and/or hydrophobicity changes of its amino acid residues. This may be due to tight structural constraints inside each proteolipid monomer or that this face of helix I11 is important for the formation of the proteolipid oligomers (52). It is worth noting that the effect of substituting these glycine residues was much more deleterious than substitutions in helix IV in the vicinity of GIu '~~ . A second site suppressor for the mutation GlylO' to Val was identified. This suppressor resulted from substituting Ile134 to Val which is located in close proximity to GIU'~~. This provides additional support to the notion that the size of the amino acid residues and not their hydrophobicity is the determination factor in the inactivation of the proteolipid and in the suppression of the inactive mutations.

Second Site Suppressc

REFERENCES 1. Mellman, I., Fuchs, R., and Helenius, A. (1986) Annu. Rev. Biochem. 56,

2. Nelson, N. (1989) J. Bioenerg. Biomembr. 21,553-571 3. Nelson, N. (1992) Biochim. Biophys. Acta 1100, 10$?-124 4. Nelson, N. (1992) Curr. Opin. Cell B id . 4, 654-660 5. Forgac, M. (1989) Physiol. Rev. SS, 765-796 6. Bowman, B. J., Dschida, W. J., Harris, T., and Bowman, E. J. (1989) J. Biol.

7. Dschida, W. J., and Bowman, B. J. (1992) J. Biol. Chem. 267, 18783-18789 8. Moriyama, Y., and Nelson, N. (1989) J. Biol. Chem. 204, 35774582 9. Stevens, T. H. (1992) J. Exp. Biol. 172, 47-55

663-700

Chem. 284,15606-15612

10. Kane, P. M. (1992) J. Exp. Biol. 172,93-103 11. Anraku, Y., Hirata, R., Wada, Y., and Ohya, Y. (1992) J. Erp. Biol. 172,6741

13. Arai, H., Berne, M., and Forgac, M. (1987) J. Biol. Chem. 262, 11006-11011 12. Sutton, R., and Apps, D. K. (1981) FEBS Lett. 130, 103-106

14. Mandel, M., Moriyama, Y., Hulmes, J. D., Pan, Y.-C. E., Nelson, H., and

15. Nelson, H., and Nelson, N. (1989) FEBS Lett. 247, 147-153 16. Lai, S., Watson, J. C., Hansen, J. N., and Sze, H. (1991) J. Biol. Chern. 266,

18. Nelson, N. (1992) Bioenerg. Biomembr. 24, 407-414 17. Sze, H., Ward, J. M., Lai, S., and Perera, I. (1992) J. Exp. Bid. 172,123-135

19. Finbow, M. E., Eliopoulos, E. E., Jackson, P. J., Keen, J. N., Meagher, L., Thompson, P., Jones, P., and Findlay, J. B. C. (1991) Protein Eng. 5, 7-15

21. Hanada, H., Hasebe, M., Moriyama, Y, Maeda, M., and Futai, M. (1991) 20. Nelson, H., and Nelson, N. (1990) Proc. Natl. Acad. Sci. U. S . A. 87,3503-3507

22. Foury, F. (1990) J. Biol. Chem. 265,18554-18560 23. Manolson, M. F., Proteau, D., Preston, R. A,, Stenbit, A,, Roberts, B. T., Hoyt,

M. A., Preuss, D., Mulholland, J., Botstein, D., and Jones, E. W. (1992) J.

24. Beltrtin, C., Kopecky, J., Pan, Y.-C. E., Nelson, H., and Nelson, N. (1992) J. Biol. Chem. 267, 14294-14303

25. Manolson, M. F., Wu, B., Proteau, D., Taillon, B. E., Roberts, B. T., Hoyt, M. A,, Biol. Chem. 267,774-779

26. Umemoto, N., Yoshihisa, T., Hirata, R., and Anraku, Y. (1990) J . Biol. Chm.

27. Noumi, T., Beltrh, C., Nelson, A,, and Nelson, N. (1991)Proc. Natl. Acad. Sci.

Nelson, N. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 5521-5524

16078-16084

Biochem. Biophys. Res. Cornmun. 176, 1062-1067

and Jones, E. W. (1994) J. Biol. Chem. 269, 14064-14074

265,18447-18453

U. S. A. 88.1938-1942 28. Ito,~ H., Fukuda, Y., Murata, R, and Kimura, A. (1983) J . Bacterial. 163,

-, ~~ ~ ~ - - -

163-168

w-s in Yeast V-ATPase 26485

29. Klionsky, D. J., Nelson, H., and Nelson, N. (1992) J. Biol. Chem. 267, 3416 3422

30. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Second Ed., Cold Spring Harbor Laboratory, Cold

31. Uchida, E., Ohsumi, Y., and Anraku, Y. (1985) J. Biol. Chem. 280,1090-1095 Spring Harbor, NY

32. Moriyama, Y., and Nelson, N. (1987) J. Biol. Chem. 262,9175-9180 33. Nelson, N. (1991) Bends Pharmacol. Sci. 12, 71-75 34. Iida, H., Yagawa, Y., a n d h a k u , Y. (1990) J. Bwl. Chem. 265,13391-13399 35. Anraku, Y., Hirata, R., Umemoto, N., and Ohya, Y. (1991) in New Era oj

Bioenergetics (Mukohata, Y., ed) pp. 133-168, Academic Press, Inc., Neu York

36. Anraku, Y., Ohya, Y., and Hidetoshi, I. (1991) Biochim. Biophys. Acta 1093, 169-177

37. Nakajima-Shimada, J., Iida, H., Tsuji, F, I., and Anraku, Y. (1991) Proc. Nutl. Acad. Sci. U. S. A. 88,6878-6882

38. Ohya, Y., Umemoto, N., Tanida, I., Ohta, A., Iida, H., and Anraku, Y. (1991) J. Biol. Chem. 266, 13971-13977

40. Tzagoloff, A., and Dieckmann, C. L. (1990) Microbiol. Rev. 64, 211-225 39. Tzagoloff, A., and Myers, A. M. (1986) Annu. Rev. Biochem. 56,249-285

41. Miller, M. J., Oldenburg, M., and Fillingame, R. H. (1990) Pmc. Natl. Acad.

42. Girvin, M. E., and Fillingame, R. H. (1993) Biochemistry 32, 12167-12177 43. Girvin, M. E., and Fillingame, R. H. (1994) Biochemistry 33,665-674 44. Fraga, D., Hermolin, J., and Fillingame, R. H. (1994) J. Bid. Chem. 269,

45. Taiz, L., and Zeiger, E. (1991)Plant Physiology, BenjamidCummings Publish.

46. Zhang, J., Myers, M., and Forgac, M. (1992) J. Bid. Chem. 267, 9773-9778 47. Jackson, S. E., Moracci, M., emastry, N., Johnson, C. M., and Fersht, A. R.

48. Supek, F., Supekova, L., Beltran, C., Nelson, H., and Nelson, N. (1992) Ann.

49. Lee, J.-I., Hwang, P. P., Hansen, C., and Wilson, T. H. (1992) J. Bid. Chem.

50. Beltrh, C., and Nelson, N. (1992) Acta Physiol. Scand. 146, 4 1 4 7 51. Boyer, P. D. (1993) Biochim. Biophys. Acta 1140, 215-250 52. Jones, P. C., Hamson, M. A., Kim, Y.-I., Finbow, M. E., and Findlay, J. B. C.

53. Lemmon, M. A., Flanagan, J. M., Treutlein, H. R., Zhang, J., and Engelman, D.

Sci. U. S. A. 87, 4900-4904

2562-2567

ing Co., Inc., Redwood City, CA

(1993) Biochemistry 32, 11259-11269

N. Y: Acad. Sci. 671,284-292

267,20758-20764

(1994) Biochem. SOC. Duns., in press

M. (1992) Biochemistry 31, 12719-12725

the journal of chemistry vol. 269, no. 42, issue of 21, pp

Documents